- Enable recursive GC marking with a limited recursion count (this option is configurable)
- No need to decrease the reference count of the gray objects anymore
- Bound function object marking is seperated into a helper function
JerryScript-DCO-1.0-Signed-off-by: Robert Fancsik frobert@inf.u-szeged.hu
This patch unifies the recursion limit checking for RegExp, function call and JSON as well.
Until now the limit was only a counter which was increased/decreased at certain points.
This counter has been substituted with a numeric limit which allows to restrict the stack usage.
This patch fixes#2963 and resolves the closed#2258 issue.
Co-authored-by: Gabor Loki loki@inf.u-szeged.hu
JerryScript-DCO-1.0-Signed-off-by: Robert Fancsik frobert@inf.u-szeged.hu
This PR is a general cleanup for garbage collection and memory
allocation code paths.
Changes:
* Removed an unnecesary local variable from 'ecma_gc_mark'.
* Refactored 'ecma_gc_run' to have an implicit list head during
iteration, which results in one less condition in the loops,
and changed the loops to use compressed pointers to reduce the
overall amount of compression/decompression.
* Renamed 'jmem_free_unused_memory_severity_t' to 'jmem_pressure_t',
and added additional values.
* Removed 'jmem_free_unused_memory_callback', instead
'ecma_free_unused_memory' is now called directly.
* Reworked 'ecma_free_unused_memory' to handle all code paths related
to 'jmem_pressure_t', and moved all relevant code paths into this
function. This simplifies the code paths in other places.
* Reworked 'jmem_heap_gc_and_alloc_block' to be more streamlined.
* Changed mem-stats to not report unused pool chunks as allocated
memory.
* Created an allocator internal API for allocating/freeing memory blocks
that are not reported as used memory in mem-stats.
* Removed iteration statistics for the jerry allocator from mem-stats,
as they don't provide any actually useful information.
Co-authored-by: Marko Fabo <mfabo@inf.u-szeged.hu>
JerryScript-DCO-1.0-Signed-off-by: Dániel Bátyai dbatyai@inf.u-szeged.hu
VM_RECURSION_LIMIT only prevented the recursion of interpreted codeblocks but
native/builtin function calls can also create stack overflow due to the too deep recursion.
This patch fixes#2905.
Co-authored-by: Gabor Loki loki@inf.u-szeged.hu
JerryScript-DCO-1.0-Signed-off-by: Robert Fancsik frobert@inf.u-szeged.hu
There are quite a few configuration macros in the project.
As discussed in the #2520 issue there are a few awkward constructs.
Main changes:
* The following macros are now 0/1 switches:
** Renamed CONFIG_ECMA_LCACHE_DISABLE to JERRY_LCACHE.
** Renamed CONFIG_ECMA_PROPERTY_HASHMAP_DISABLE to JERRY_PROPERTY_HASHMAP.
** Renamed CONFIG_DISABLE_UNICODE_CASE_CONVERSION to JERRY_UNICODE_CASE_CONVERSION.
** Renamed ENABLE_REGEXP_STRICT_MODE to JERRY_REGEXP_STRICT_MODE.
** Renamed JERRY_DISABLE_JS_PARSER to JERRY_PARSER.
** Renamed JERRY_ENABLE_ERROR_MESSAGES to JERRY_ERROR_MESSAGES.
** Renamed JERRY_ENABLE_EXTERNAL_CONTEXT to JERRY_EXTERNAL_CONTEXT.
** Renamed JERRY_ENABLE_LINE_INFO to JERRY_LINE_INFO.
** Renamed JERRY_ENABLE_LOGGING to JERRY_LOGGING.
** Renamed JERRY_ENABLE_SNAPSHOT_EXEC to JERRY_SNAPSHOT_EXEC.
** Renamed JERRY_ENABLE_SNAPSHOT_SAVE to JERRY_SNAPSHOT_SAVE.
** Renamed JERRY_SYSTEM_ALLOCATOR to JERRY_SYSTEM_ALLOCATOR.
** Renamed JERRY_VM_EXEC_STOP to JERRY_VM_EXEC_STOP.
** Renamed JMEM_GC_BEFORE_EACH_ALLOC to JERRY_MEM_GC_BEFORE_EACH_ALLOC.
** Renamed JMEM_STATS to JERRY_MEM_STATS.
** Renamed PARSER_DUMP_BYTE_CODE to JERRY_PARSER_DUMP_BYTE_CODE.
** Renamed REGEXP_DUMP_BYTE_CODE to JERRY_REGEXP_DUMP_BYTE_CODE.
* Recursion check changes:
** Renamed REGEXP_RECURSION_LIMIT to JERRY_REGEXP_RECURSION_LIMIT.
** Renamed VM_RECURSION_LIMIT to JERRY_VM_RECURSION_LIMIT.
* Attribute macro changes:
** Renamed JERRY_CONST_DATA to JERRY_ATTR_CONST_DATA.
** Renamed JERRY_HEAP_SECTION_ATTR to JERRY_ATTR_GLOBAL_HEAP.
Now the macro can specify any attribute for the global heap object.
* Other macro changes:
** Renamed CONFIG_MEM_HEAP_AREA_SIZE to JERRY_GLOBAL_HEAP_SIZE.
Then new macro now specify the global heap size in kilobytes.
* Updated documentations to reflect the new macro names.
For more deatils please see jerry-core/config.h.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal pgal.u-szeged@partner.samsung.com
There are quite a few configuration macros in the project.
As discussed in the #2520 issue there are a few awkward constructs.
Main changes:
* Renamed all CONFIG_DISABLE_<name>_BUILTIN macro to JERRY_BUILTIN_<name> format.
* The special JERRY_BUILTINS macro specifies the basic config for all es5.1 builtins.
* Renamed all CONFIG_DISABLE_ES2015_<name> to JERRY_ES2015_<name> format.
* The special JERRY_ES2015 macro specifies the basic config for all es2015 builtins.
* Renamed UNICODE_CASE_CONVERSION to JERRY_UNICODE_CASE_CONVERSION.
* Renamed ENABLE_REGEXP_STRICT_MODE to JERRY_REGEXP_STRICT_MODE.
* All options (in this change) can have a value of 0 or 1.
* Renamed ENABLE_REGEXP_STRICT_MODE to JERRY_REGEXP_STRICT_MODE.
JERRY_REGEXP_STRICT_MODE is set to 0 by default.
* Reworked CONFIG_ECMA_NUMBER_TYPE macro to JERRY_NUMBER_TYPE_FLOAT64 name and now
it uses the value 1 for 64 bit floating point numbers and 0 for 32 bit floating point
number.
By default the 64-bit floating point number mode is enabled.
* All new JERRY_ defines can be used wit the `#if ENABLED (JERRY_...)` construct to
test if the feature is enabled or not.
* Added/replaced a few config.h includes to correctly propagate the macro values.
* Added sanity checks for each macro to avoid incorrectly set values.
* Updated profile documentation.
* The CMake feature names are not updated at this point.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal pgal.u-szeged@partner.samsung.com
This patch adds posibility to supervise the VM call stack to avoid aborts/crashes due to the recursion calls.
JerryScript-DCO-1.0-Signed-off-by: Robert Fancsik frobert@inf.u-szeged.hu
It is superfluous to maintain multiple globals when the whole
reason of context is to keep them in a single place. It also
simplifies initialization and external context creation a bit.
Also removed unused lcache header includes.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss akiss@inf.u-szeged.hu
Aborts are similar to exceptions except they are not caught by catch
and finally blocks. Callbacks should honor aborts as well and return
them without processing them. Aborts are never thrown by JavaScript
code.
In the future certain events such as out-of-memory condition may
also throw aborts.
JerryScript-DCO-1.0-Signed-off-by: Zoltan Herczeg zherczeg.u-szeged@partner.samsung.com
* Move the job queue from the ports into jerry-core
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss akiss@inf.u-szeged.hu
* Remove port notification and keep `jerry_run_all_enqueued_jobs` API only
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss akiss@inf.u-szeged.hu
* Introduced new 'jmem.h' for public part.
* Removed obsolete headers
JerryScript-DCO-1.0-Signed-off-by: László Langó llango.u-szeged@partner.samsung.com
Since the project is now hosted at the JS Foundation we can move to unified copyright notices for the project.
Starting with this commit all future contributions to the project should only carry the following copyright notice (except for third-party code which requires copyright information to be preserved):
"Copyright JS Foundation and other contributors, http://js.foundation" (without the quotes)
This avoids cluttering the codebase with contributor-specific copyright notices which have a higher maintenance overhead and tend to get outdated quickly. Also dropping the year from the copyright notices helps to avoid yearly code changes just to update the copyright notices.
Note that each contributor still retains full copyright ownership of his/her contributions and the respective authorship is tracked very accurately via Git.
JerryScript-DCO-1.0-Signed-off-by: Tilmann Scheller t.scheller@samsung.com
The algorithm is the following:
* introduced a counter variable, which value can be [0-4]
* if its value is 0, property hashmap allocation is enabled
* JMEM_FREE_UNUSED_MEMORY_SEVERITY_LOW -> decrease the counter
* JMEM_FREE_UNUSED_MEMORY_SEVERITY_HIGH -> increase the number
* if JMEM_FREE_UNUSED_MEMORY_SEVERITY_HIGH happens twice in a row increase the counter to 4
According to the measurements this algorithm provides better runtime results in low memory conditions.
JerryScript-DCO-1.0-Signed-off-by: István Kádár ikadar@inf.u-szeged.hu
Zero out all globals (and remove unnecessary init() functions).
Move snapshot globals to a temporary stack variable.
JerryScript-DCO-1.0-Signed-off-by: Zoltan Herczeg zherczeg.u-szeged@partner.samsung.com
- New allocator is added that returns null on out of memory, property hasmap create uses this allocator for now.
- Property hashmaps of objects are removed durring a high severity gc.
Follow up patch is in progress.
JerryScript-DCO-1.0-Signed-off-by: István Kádár ikadar@inf.u-szeged.hu
The new literal storage keeps ecma strings rather than having a
custom string implementation which duplicates the string management
routines. Conversions between string implementations are eliminated
which improved the performance by 4%.
JerryScript-DCO-1.0-Signed-off-by: Zoltan Herczeg zherczeg.u-szeged@partner.samsung.com
The cache stores only real properties now, because storing NULLs has
little benefit according to tests. Since only real properties are
stored now, there is no need to create real references to objects
and property names, which reduces the keeping of dead objects after
garbage collection.
JerryScript-DCO-1.0-Signed-off-by: Zoltan Herczeg zherczeg.u-szeged@partner.samsung.com
The 'mem_' prefix is too general, so it might clash with
symbols in other libraries. Renamed the directory, file,
funtion and type names.
Related issue: #1052
JerryScript-DCO-1.0-Signed-off-by: László Langó llango.u-szeged@partner.samsung.com
